Lung ventilators provide therapeutic gas (e.g., oxygen) and pressure volume support to a patient's lungs in order to facilitate gas exchange with a patient, either by supporting spontaneous breathing or by providing mandatory ventilation in the absence of spontaneous breathing. The gas is provided to a patient through an inspiratory conduit and the ventilator is fluidly coupled to the lung using a patient interface such as an endotracheal tube, a nasal cannula or a mask. There is a risk that fluid connection from the ventilator to the lung may be lost, for example due to movement of the patient causing the patient interface to become dislodged. If the patient interface is disconnected from the patient, this loss of breathing can be harmful to the patient.
Current ventilators detect such loss of breathing by use of complex algorithms based on measurements of breathing circuit pressure and flow. These algorithms can be problematic in scenarios with significant leaks, such as are common in long term applications, or with high-resistance tubes such as narrow bore nasal cannulae. When a breathing tube is dislodged, it may remain occluded or partially occluded, for example, by resting against the patient's face, causing the alarm mechanism to fail. Still further, these complex algorithms can be difficult to implement within the ventilator. As a result, reliable detection methods and adequate sensitivity to breathing circuit integrity are desired.